top of page

Joke Wars : Robot vs Human

Do you think you have a nice sense of humour? And do you think a robot can roast you with jokes? Let’s give that a look-see.

How do robots pay for things?

With cache, of course!

This style of joke-telling is the perfect starting point for the extravagant quest to develop robots that have a sense of humour (lower than mine, of course, not everybody can reach my limits ;)).

I’m not talking about the "robot fail" videos you see on social media (though they’re intensely funny sometimes). AI scholars are trying to develop robots that are intentionally funny. These robots can detect the type of joke you told and respond with an equally amusing witty response.

Some AI experts believe that joke-telling robots will be the final word on robots' being identical to humans. Instilling humour in a robot requires a strong grasp of complicated aspects of being human, like empathy, self-awareness, and being expressive about your emotions.

We aren’t quite there yet. Let’s seek the answer to the burning question: "Joke-telling robots: Will this be the final step required for a robot to be exactly identical to a human?"

Will this be the endgame?

Julia Taylor, a professor at Purdue Polytechnic and an expert on computational humour, stated "From the psychological point of view, we are not quite sure what it takes," she said. What kind of sense of humour will people have? What is appropriate and what is not appropriate?

"All of those difficulties are there before you can actually get AI-type humour."

AI knows only the information that humans choose to provide it with. This means that if we want a robot to make us laugh, we have to be as crystal clear as stars on a frosty night about the type of humour we want to train it with.

Humour is variably subjective. Everyone likes different types of humour. I find jokes that you wouldn't find funny, and that’s the exact reason why joke-telling robots can never be perfect.

To understand and evaluate each person’s level and type of humour, these robots will have to have a lot more work done on them (like the Kardashians have already done ;) ). There also arises a situation where our own tastes in humour may change from time to time. Maybe you’ll appreciate a joke now, but after 2-3 days, maybe you won't appreciate it anymore.

So how do we instruct a robot to give and collect laughter if we don't quite understand what we find funny and why?

"When humans find something funny, we may not know why it is funny," Taylor said. "We are working on the theory of humour, trying to define all of its subsets of it. We're not quite there. There are a lot of methodologies towards theories, but most people will agree that those theories are not fine-grained enough to be implemented by a computer. Not yet. "

This means that for a while, computational humour is restricted to the areas that it is comfortable in. So, for now, humour remains a people thing exclusively. If we want to have computers truly understand jokes, somehow come up with their own and know when and to whom to tell them, we’re going to have to upload into them the entirety of humanity. There are already many other joke-telling robots out there on the market, some of which crack actually funny jokes once in a blue moon. Let’s tour the utopia of painfully (un) funny robots!

  1. DEviaNT stands for Double Entendre via Noun Transfer. Cracks "that's what she said" jokes. According to its developers, 72% of the time, DEviaNT gets it right, and "says" "that's what she said" after a serious sentence, and it's too funny for words. Steve Carrell is known for cracking these kinds of jokes in ‘The Office’. Do you think this robot could possibly replace him and maybe even crack better jokes? I believe that this could plausibly be a big step forward.

  2. Semi-supervised Algorithm for Sarcasm Identification (SASI) This machine learning algorithm was designed to help AI by recognise sarcasm. They currently report a 77% success rate. Scientists are realising that the detection of sarcasm is a very important and useful tool for humans and would certainly mean an advancement in AI that is commendable.

  3. STAND-UP is a system that uses puns to augment non-verbal dialogue. This programme was created by a team of researchers to assist children who use computerised speech aids to help them with certain challenges faced by them during communication.

  4. Manatee This is the joke-generation system developed by professor Kristian Hammond and PhD student Patrick McNally. To come up with its jokes, Manatee searches the Internet for word combinations that fit into widely known witty phrases, such as "I like my X like I like my Y." Then it incorporates the result into a related photo to create a webcomic, something like this:,c_limit/joke-1.jpg

Should we be relieved or concerned?

All of these attempts sum up to this point: up until now, computers have been able to tell jokes, but only absolutely moronic ones.

Check out this computer-generated quip:

What is the difference between leaves and a car?

One brushes and rakes, the other rushes and brakes.

If it ever reaches a point where robots have taken over the world, just know that we’ll be in for a bleak future with really horrible wit.

Jokester is one of Dr Isaac Asimov's gems, which is a delightful short story on machine humour. I definitely recommend it.

I'd like to leave you with a thought: do you think this computational humour will ever reach the point where it can cause harm?


bottom of page